Temporal and spatial changes and driving mechanisms of the conversion from paddy fields to drylands in China

Arable land is essential for food security and sustainable socio-economic development. In China, extensive areas of paddy fields have been converted to drylands due to natural, socio-economic, and policy pressures, yet the underlying mechanisms remain insufficiently understood. Using multi-temporal land use data (1980–2020), ArcGIS, and the geographical detector, this study investigates the spatiotemporal dynamics and drivers of such conversion. The converted area expanded from 5,386 km 2 during 1980–1990 to 27,147 km 2 during 2010–2020, with growth accelerating markedly after 2000, nearly doubling the rate of the previous two decades. Spatially, the pattern shifted from concentration in the low-elevation, low-slope zones of Northeast China to dispersion into central and southwestern mountainous regions, revealing a distinct topographic gradient. Ecosystem service values (ESV) increased overall (+621 Million CNY), but regulating services declined (–16,426 Million CNY), indicating trade-offs. Precipitation reduction and elevation emerged as dominant natural drivers, while labor shortages and depopulation were key socio-economic factors; their interactions exhibited nonlinear enhancement. This study integrates multi-temporal spatial analysis with quantitative driver detection, underscoring the influence of topography on conversion. Policy recommendations include strictly maintaining the farmland “red line,” enhancing soil and water management to reduce ecological losses, and promoting high-value dryland crops to secure both food production and rural livelihoods. These findings provide scientific evidence for sustainable arable land management and rural revitalization in China. • Significant temporal and spatial shifts occurred in the conversion of paddy fields to drylands in China from 1980 to 2020, with a marked increase and pattern dispersion after 2000. • The conversion exhibited clear topographical gradients and led to an overall rise in ecosystem service value, but notably reduced regulating services. • Natural, locational, and socio-economic factors jointly drove the conversion, with geographical detectors revealing stage-specific and synergistic effects.